WHY NICARAGUAN PEASANTS REMAIN IN AGRICULTURAL PRODUCTION COOPERATIVES
Zvi Lerman Department of Agricultural Economics and Management, The Hebrew University of Jerusalem, Israel lerman@agri.huji.ac.il
Ruerd Ruben Development Economics Group, Wageningen University, The Netherlands Ruerd.Ruben@wur.nl
Selected Paper prepared for presentation at the American Agricultural Economics Association Annual Meeting, Providence, Rhode Island, July 24-27, 2005
Copyright 2005 by Zvi Lerman and Natalya Shagaida. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies.
Why Nicaraguan Peasants Remain in Agricultural Production Cooperatives1
Zvi Lerman and Ruerd Ruben Abstract A substantial number of Nicaraguan peasants remain in agricultural production cooperatives created as part of the Sandinista land reform in the 1980s. This phenomenon persists despite the current regime’s support for parcellation of the land-reform cooperatives and contrary to theoretical considerations that regard individual farms as more productive and more efficient. We aim to identify the factors that govern the choice of Nicaraguan peasants between staying in a cooperative and exiting to start an individual farm. Sample data from 475 landed households in four agro-ecological macro-regions are used to compare resources, incomes, and productivity of peasants remaining in production cooperatives, former members of cooperatives who became independent farmers, and peasants who had always farmed independently. Low capital endowments of the new independent farmers adversely affect their standard of living. Better access to non-farm income encourages peasants to remain in the cooperative, where they are less exposed to risk. Uncertainty associated with land ownership rights and difficulties with resolution of cooperative debt play a dominant role in keeping Nicaraguan peasants in cooperatives. Cooperative members maintain a high degree of coordination of activities and are thus able to capture the benefits of rural development programs. Keywords: cooperative farms, individual farming, land reform, Latin America, social capital. Introduction: Persistence of Cooperatives The cooperative experience in Nicaragua followed the web and flow of domestic politics. During the four decades of the Somoza regime (1936-1979), the structure of agricultural holdings in Nicaragua had an extremely polarized distribution consisting of a small number of extensive
1
private estates and a large number of very small family plots. The political and social failures of the Somoza regime led to the Sandinista revolution in 1979, and the new ruling party implemented its socialist agenda by expropriating the land of the large estate owners and allocating it to the use of the rural landless. However, the peasants were encouraged to join agricultural production cooperatives (APCs) rather than use the allocated plots for individual farming. The Sandinista policies thus led to the creation of large agricultural cooperatives in place of former privately owned estates and did relatively little to enlarge and strengthen the sector of family farms (Baumeister 1998; Everingham 2001). In the best socialist tradition, cooperatives enjoyed generous support under the Sandinista government. After about a decade, the socialist-oriented Sandinista regime was replaced in democratic elections by a market-oriented government. The new government policies emphasized liberalization of the entire economy and abandoned the former preferential treatment of agricultural cooperatives. The change in government policies and attitudes led to a wave of liquidations of agricultural cooperatives, and many peasants left their cooperatives with a plot of land for individual farming. Yet not all the agricultural production cooperatives were liquidated despite the changes in the political atmosphere, and some continue to exist to the present day. The persistence of agricultural production cooperatives in Nicaragua raises an interesting research question. Theoretical considerations and empirical evidence suggest that individual farms are more productive and more efficient than agricultural production cooperatives (see, e.g., Hazell 2003). Why is it then that the Nicaraguan cooperatives did not fully split up into individual farms once the government had abolished the policy constraints favoring cooperatives? Is it possible to identify the factors that govern the choice of Nicaraguan farmers between remaining in a cooperative and exiting to start an individual farm?
2
A similar change in the policy environment took place in the transition economies in the former Soviet Union and Central Eastern Europe, where government policies dramatically shifted from unquestioning support of collective farming as a mandatory organizational form in agriculture to a more liberal market-oriented attitude that allowed individual farming (Brooks 1993). The initial expectations were of a rapid breakup of large-scale collective farms2 and a sweeping transition to individual farming on land withdrawn from the former collectives. Yet these expectations have not materialized so far and large-scale collective structures continue to exist in all former socialist countries (although now they are organized as large-scale corporations that cultivate land not withdrawn for individual farming). The persistence of collective (or corporate) farming in transition countries – contrary to the standard pattern of agriculture in market economies – is an intriguing topic that attracts considerable attention among scholars (Amelina 2000; Deininger 1995). Unfortunately, no satisfactory analytical results are available to explain this phenomenon in transition countries beyond the usual argument that peasants try to avoid new risks by staying under the familiar collective umbrella. Nicaraguan data could shed some light on the reasons for the persistence of cooperative farms in environments where policies purportedly encourage individual farming. This study is based on a representative survey of 475 landed households in four agroecological regions located in the central part of Nicaragua, where agriculture is an important component of rural livelihoods.3 These macro-regions cover various agro-ecological conditions and farming systems that are typical for the Nicaraguan countryside. The sample was designed to include three distinct groups of respondents: (1) peasants who continued to farm in agricultural production cooperatives, (2) former members of agricultural production cooperatives who decided to take up independent farming, and (3) peasants who had always been engaged in
3
independent farming.4 The survey was carried out in 2000 and focused on the analysis of underlying differences in farm household characteristics, level and composition of household income, assets and wealth, and efficiency of production systems as possible reasons for staying in or exiting from the cooperative framework.5 The remainder of the article is structured as follows. We provide a short overview of land policies that led to the formation of agricultural production cooperatives in Nicaragua during the 1980s and subsequent programs for individual land titling during the 1990s. Main factors that influence the socio-economic position of cooperative members vis-à-vis old and new independent peasants are highlighted. Then follows an empirical analysis of the income position and living standards of each of the three groups of peasant households. We address possible differences in production efficiency among the three groups, concluding that these are hardly sufficient to explain any inherent superiority of either individual or cooperative farming. We accordingly proceed to analyze the role of other internal and external factors that influence the decision to stay in the cooperative or to become an independent peasant. Attention is focused on differences in human, physical, and social capital that determine individual preferences regarding the desired institutional arrangements. We conclude with some policy suggestions for removing the constraints that still prevent Nicaraguan peasants from making a free choice between alternative organizational forms of independent farming and cooperative membership. These suggestions are guided by the general view that minimizing the constraints on the activities of economic agents is conducive to economic efficiency. Rural Organization in Nicaragua Nicaragua provides a challenging environment for the analysis of changing patterns of rural organization. During the last decades, the roles and functions performed by state, market, and
4
community organizations have been frequently modified. The former Sandinista government (1979-1989) placed great emphasis on the role of state agencies for trade, credit, and extension services. The land reform program favored – after an initial preference for state farms – the establishment of a large number of production cooperatives in the agrarian sector.6 The rural cooperative sector in Nicaragua represented in 1989 about 21% of agricultural land use, 20% of the rural population, and 24% of agrarian production (San Martin 1992).7 Subsequent liberal regimes strongly reduced credit support and service delivery to cooperatives and permitted the parcellation of cooperative land, leading to a distinct decline in cooperative membership. At the end of 1999 the rural cooperative sector was reduced to no more than 9% of the land and 8% of the rural population (Ruben and Masset 2003). The partial disintegration of the land-reform cooperatives gave rise to a number of diverse pathways of organizational change. Some peasants left the cooperative and took up individual farming, while others decided to remain cooperative members. The process of land reform during the Sandinista regime emphasized the creation of agricultural production cooperatives as an alternative for poor and often landless households to get access to land, credit, and extension services. Collective ownership was initially promoted as a device to maintain rural stability and to allow effective delivery of public services. Peasant organizations also readily embraced the cooperative mode due to the perceived advantages of economies of scale and as a means to reinforce their negotiating position vis-à-vis the government. However, internal organization of newly established APCs was largely based on close interaction between peasant households that maintain strong family ties (Carter el al. 1993). Collective fields were usually reserved for cash crop production while members maintained individually operated subsidiary plots to satisfy their home consumption needs. Labor services supplied to operations in
5
collective fields were paid for on a weekly or monthly basis and represented an advance payment against the harvest. Most Nicaraguan APCs therefore relied on bank credit to purchase farm inputs and to pay for members’ labor. Peasants’ willingness to engage in cooperative production has traditionally been attributed to motives of income generation and risk sharing. Several internal and external problems make farmers averse to full participation in collective resource management. Free-riding behavior and non-compliance are frequently mentioned in the literature as motives for leaving the cooperative. In addition, policies towards market liberalization, financial reforms, and new legislation regarding ownership encourage the parcellation of land-reform cooperatives. Yet legal uncertainties regarding land ownership, unclear entitlements of individual members to collectively owned assets, and difficulties with the resolution of outstanding debts inhibit in practice any straightforward subdivision. Therefore, a wide range of institutional arrangements have emerged, including peasants who stick to APC membership and others who prefer independent production or seek loose affiliation that still guarantees access to some services. The disadvantages of large-scale, capital-intensive cooperative farms are generally acknowledged. Well-known problems of under-investment, labor shirking, productivity decline, and membership desertion affect cooperative performance.8 If access to bank credit is substantially reduced, cooperatives are likely to disintegrate as they are inherently unable to raise equity or private loans and are thus left without financial sources for paying to members engaged in collective activities. The promulgation of a new legal framework that encourages the issuing of individual titles or landownership certificates further reinforces this tendency towards parcellation.
6
Modern approaches to rural organization focus on aspects of risk sharing, access to information, and reduction of transaction costs as potential reasons for maintaining cooperative relations (Hoff et al. 1993; Bardhan 1989). Common-property resource management is also seen as a device to control natural resource depletion (Bromley 1992; Wilson and Thompson 1993). It has been demonstrated that resource-use efficiency on collective parcels is not necessarily inferior to that on private parcels (Ruben 1999). The importance of cooperative ties for securing peasants’ livelihoods in rural Nicaragua is widely acknowledged (Jonakin 1992; Carter et al. 1993). Medium-size agricultural cooperatives that rely on linkages between semi-independent family plots could therefore still be attractive to peasant households (Carter 1987; Putterman 1981; Hussi et al. 1993). Earlier research on agrarian cooperatives is related to large-scale (mechanized) farms under socialist-oriented political regimes. In Nicaragua, most APCs are medium-size enterprises with labor-intensive production systems. These cooperatives were originally established during the land-reform process, and it is interesting to analyze their behavior when market-oriented policies are put in place. Contrary to expectations triggered by the theoretical disadvantages of cooperative organization, APC members proved fairly reluctant to proceed towards full parcellation of land to individuals. The decision to continue APC membership or switch to independent farming can be viewed as an endogenous process where various internal and external factors interact. Indeed, Nicaraguan peasants have to face institutional choice at two distinct levels: (1) the decision whether to stay in the cooperative or switch to independent farming, and (2) the decision whether to rely on the cooperative as a channel for getting access to credit, services, and information or to rely on individual interaction with external agents. The first aspect refers to the existing opportunities for improving the members’ socio-economic position through
7
reallocation of available human and physical capital, while the second aspect relates to the possibilities for reliance on social capital as a device to guarantee access to inputs and output markets and thus improve household welfare. Even if independent farming enables peasants to improve resource management, limited access to markets and institutions may easily reduce this potential advantage to nil. Endowments, Incomes, and Productivity We analyzed the socio-economic situation of peasants who had always been independent producers (group I, labeled as “private”), peasants who had recently left the cooperative to take up independent farming (group II, labeled as “parcellation”), and households that continued cooperative membership (group III, labeled as “cooperation”). Subdivision of cooperative farms is only expected to be attractive when household income or farm productivity can be increased. Theoretical considerations suggest a prior ranking where established private farmers perform better than newly “parceled” farms, whereas the latter outperform cooperative farms (i.e. group I > group II > group III). We used standard ANOVA tools to identify significant differences among the three groups of producers in terms of resource endowments, income profiles, and living standards. In addition, estimated production functions were used to analyze differences in factor productivity. The household profiles presented in Table 1 do not show a consistent pattern of differences across the three groups of peasants. To the extent that pairwise comparisons detect any differences between groups, group I (private farmers who have always remained independent) deviates most from the other two groups. No significant differences by any of the variables are observed between group II (“parcellation” farmers who have left the cooperative to take up independent farming) and group III (farmers who remain cooperative member). An
8
important conclusion from Table 1 is that group I farmers have more land, more livestock, and more equipment than group II farmers (only the differences in land and livestock are statistically significant). The new independent farmers who have left their cooperative are less endowed with basic physical capital compared to established private farmers. These differences in physical capital are reflected in higher levels of total family income and especially farm income for group I farmers compared with the other two groups. Moreover, the mean farm income for the new independent farmers in group II seems to be higher than for group III farmers remaining in cooperatives. The farm income results are thus consistent with the hypothesis formulated above. Unfortunately, these income differences are not statistically significant (due to the high variability in the sample data) and thus cannot be taken as real support for our hypothesis.
Table 1. Profiles of farm households in the survey (averages per household) Whole Group I: Group II: Group III: Pairwise sample “Private” “Parcellation” “Cooperation” differences* (N=476) (N=155) (N=154) (N=167) Human capital Family size (persons) 6.7 6.3 7.1 7.0 I < II Children (younger than 8) 1.6 1.4 1.6 1.7 None Adults (older than 8) 5.2 4.9 5.4 5.2 None Average household age 33 36 32 31 I > II; I > III Average years of schooling 4.3 4.2 4.4 4.5 None Physical capital Land (manzanas)1 21.3 24.7 18.98 20.0 I > II Livestock (number of animals) 9.3 12.6 6.5 8.2 I > II; I > III Farm equipment (pieces) 5.9 6.2 5.8 5.8 None Income Family income (cordobas)2 24,148 29,078 21,880 21,664 None Farm income (cordobas)2 16,513 20,635 16,058 13,106 None 2,3 Non-farm income(cordobas) 7,636 8,443 5,882 8,558 None Share of non-farm income3,4 50% 44% 48% 57% None * Differences shown are statistically significant at 10% by the Bonferroni multiple pairwise comparisons test. 1. Manzana is a unit of land area: 1 manzana = 0.7 hectares. 2. Income in cordobas (US$1 = 7 cordobas at the time of the survey). 3. Non-farm income includes wages received from the cooperative. 4. Simple average of non-farm income shares for each respondent.
While the survey does not reveal statistically significant differences in income across the three groups, the data nevertheless clearly show that the total family income – and especially the
9
farm component of family income – increases with the increase of the amount of land available to the family. This phenomenon is demonstrated in Figure 1. Total family income rises from 11,400 cordobas (US$1,630) for families with the smallest land plots (up to 2 manzanas, or 1.4 hectares) to more than 40,000 cordobas (US$5,700) for families with 60 manzanas (42 hectares) and up. The non-farm income remains fairly stable in absolute value across all farm-size groups, and as a result its share in household income drops from about 85% for families with the smallest plots to about 15% for families with relatively large land endowments. Consistently with the situation in many developing countries, land endowment is a major factor in improving family income and well-being.
Figure 1: Land Endowments and Famil y Income
thousand of cordobas 80
60 Farm Non-farm
40
20
0 0-2
2-5
5-10
10-20 20-40 40-60 60-80 manzanas
>80
The descriptive univariate analysis presented above has been extended to a multivariate production-function framework to detect differences in factor productivity across the three groups. Theory suggests that independent farms (groups I and II) are more productive than farms operating in a production cooperative (group III). To test this hypothesis, we estimated a
10
standard Cobb-Douglas production function regressing farm income (in cordobas) as the dependent variable on land, the number of animals, the number of farm workers, and the number of pieces of farm machinery as the explanatory variables. The model also included a dummy variable corresponding to different groups of respondents (groups I, II, and III representing established independent farmers, peasants who have left the cooperative through parcellation, and peasants who continue to work in a cooperative).
Table 2. Estimation of Cobb-Douglas production function with group dummy variables Dependent variable : Farm income Parameter estimate Significance level Explanatory variables Intercept (Group III) 6.621 0.000 Land 0.310 0.001 Animals 0.236 0.002 Labor -0.165 0.124 Machinery 0.603 0.004 Group I (“private”) -0.062 0.754 II (“parcellation”) 0.204 0.305 III (“cooperation”) 0.000 -R2 = 0.16 ; F-value = 11.07 (p = 0.000)
The estimation results are presented in Table 2. All factors (except labor) have significant positive coefficients, with farm machinery displaying the highest elasticity (0.6). The coefficient for labor is not statistically significant and has a negative sign, which may be indicative of excess labor on Nicaraguan farms. Such “overloading” is a frequently observed phenomenon in rural areas where land quality is poor and off-farm employment options are limited (Kennedy 1983). The main conclusion of the analysis, however, is that there are no clear-cut differences in productivity among the three groups: the intercept shifters (for groups I and II relative to group III) are not statistically significant in the sample. The results of these analyses do not fully verify the general theoretical hypothesis that group I > group II > group III by relevant performance measures.9 To gain additional evidence for this hypothesis, we have examined some qualitative questions included in the subjective attitude part of the questionnaire. These questions provide information about the perceived
11
standard of living of the respondent families. One particular question asks whether the family has enough availability of food. Other questions attempt to get a more detailed assessment of the standard of living by exploring what the family can buy with the available income. The responses to both questions indicate that established private farmers experience a higher standard of living: nearly 60% of group I respondents report that they have enough food – compared to less than 50% in the other two groups; for over 20% of group I respondents the family income is sufficient to buy more than just food and basic daily necessities – compared to 12% in the other two groups.
Table 3. Perceived adequacy of family income as a measure of standard of living (percent of respondents by group) Group I: Group II: Group II: “Private” “Parcellation” “Cooperation” Food security: Does the family have enough food? Yes 58 42 48 Standard of living: What does the family income buy? Lowest: Not enough even for food 37 42 39 Medium: Just enough for food and basic necessities 43 46 49 Highest: Enough to buy more than just food and basic necessities 21 12 12 Total 100 100 100
The differences in these percentages between group I (established private farmers) and groups II and III combined (former and present cooperative members) are statistically significant. The differences between groups II and III are not statistically significant. Nevertheless, the answers seem to indicate that group II respondents – the new independent farmers who have left their cooperative – have the lowest standard of living in the sample, even lower than the respondents who continue to belong to a cooperative. These qualitative results thus suggest that the standard of living ranking is group I > group III ≥ group II rather than group I > group II > group III. A possible explanation for this unexpected pattern probably can be found in the low capital endowments of group II farmers. Former cooperative members somehow do not get a fair deal when they leave the cooperative, receiving little land and
12
relatively few animals (see Table 1). These low capital endowments adversely affect their standard of living as new independent farmers. Reasons to Stay: The Role of Human and Physical Capital The better standard of living and the higher family incomes among independent farmers usually may provide an incentive for switching from cooperative to individual farming. Previous research suggests that the decision to leave an agricultural production cooperative and switch to independent farming is determined by the availability of both human and physical capital in the household (Deininger 1993; Putterman 1985). Human capital includes variables that describe the household’s labor pool (e.g., the number of adults in the household, age composition, and educational attainment). Physical capital includes land, livestock, farm machinery and buildings, and also family income, which is a source of funds for both working capital and investment, as well as a safety net for the family under adverse conditions. We would expect that a greater capital stock (both human and physical) should have a positive impact on the decision to leave the cooperative and start farming independently. People with more capital can afford to sacrifice the relative security of a cooperative and opt for the higher risks of independent farming, with its promise of commensurately higher returns. We performed a logistic regression analysis to identify the impact of human and physical capital variables on the decision to leave the cooperative for independent farming.10 This impact is expressed by the effect on the probability of switching from cooperative to independent farming. If the impact is positive, an increase of the corresponding variable increases the probability of switching from cooperative to independent farming and we accordingly expect to get a positive coefficient in the estimated logistic model. If the impact is negative, an increase of the corresponding variable decreases the probability of switching from cooperative to
13
independent farming and we accordingly expect to get a negative coefficient in the estimated logistic model. The logistic model in our analysis includes terms for age squared and years of schooling squared to capture nonlinear human capital effects, which are often discussed in the literature (Lerman et al., 2004). Regarding the effect of age, it is expected that as a person grows older, it becomes more difficult to change established patterns of behavior and older people are therefore less likely to leave the cooperative. However, the age variable spans the whole lifetime and its effect may change over the person’s life. Very young persons, without proper training and experience, are not ready to leave the cooperative. As young people become more mature and experienced, they may be willing to start a new way of life and are thus expected to exhibit a higher likelihood of leaving the cooperative than very young people. Beyond a certain age conservatism may predominate and the person will be more likely to stay in the cooperative. These two opposing effects are captured by a combination of linear and quadratic terms in the model: the willingness of young people to leave the cooperative is demonstrated by the linear term (with an expected positive sign), while the reluctance of older people to leave is reflected in the quadratic term (with an expected negative sign). The quadratic term becomes dominant as age increases, and on the whole age will have a negative impact on the probability of leaving the cooperative. Similar considerations suggest that increases in education at a very basic level (few years of schooling) will not necessarily increase the likelihood of leaving the cooperative, yet as the educational endowment increases beyond a certain basic level, people will become ready to start a new life as independent farmers. This effect is captured by a combination of linear and quadratic terms representing the education variable.
14
The income variable was separated into two components: farm income and income from non-farm sources. This was done because the theoretical interpretation of the impact of non-farm income is not clear-cut. It can be argued that non-farm income provides a cushion against risk and, similarly to the effect of farm income, increases the probability of leaving the cooperative in favor of independent farming. On the other hand, it can also be argued that non-farm income derives in part from salaries paid by the cooperative and thus encourages the family to stay. If the second interpretation is true, then non-farm income should have a negative impact on the probability of leaving the cooperative. Because of this ambiguity, farm and non-farm income were included as separate explanatory variables in the logistic regression model.
Table 4. Individual choice between staying in a cooperative and becoming an independent farmer (logistic regression analysis) Explanatory variables Parameter Significance Expected impact on the probability of estimate level switching from cooperative to independent farming Intercept -3.621 0.042 Human capital Age of head of household 0.151 0.037 Negative when age is sufficiently high; Age squared -0.001 0.068 positive for young people Education of head of household -0.104 0.332 Positive for high educational endowEducation squared 0.009 0.425 ments; zero at low level of education Number of adults in household -0.024 0.693 Positive Physical capital Land 0.002 0.787 Positive Number of animals -0.041 0.008 Positive as a wealth factor; may be negative if pasture land is constrained Farm income 9.0E-6 0.058 Positive Non-farm income -20E-6 0.102 Positive (higher ability to take risk) or negative (less-dependent on farming)
The logistic regression results based on the survey data are presented in Table 4. The table compares the signs of the estimated coefficients with the expected signs from the hypotheses discussed above. Overall, the results do not fully support the viewpoints of the human and physical capital approaches. Among the human capital variables, the impact of age is consistent with the prior hypotheses: the linear term has a positive coefficient, while the quadratic term has a negative coefficient, and both coefficients are statistically significant (at
15
10%). The effect of education is not statistically significant, although the quadratic term has a positive sign as expected. The family labor pool does not have a significant effect on the decision to switch to individual farming. The results are less ambiguous for the physical capital variables. The coefficients of land and farm income have positive signs that are consistent with our prior hypotheses, but only the coefficient of farm income is statistically significant. The coefficient of non-farm income is negative with marginal statistical significance. Its impact is thus consistent with the second interpretation offered in our discussion of the hypotheses: higher non-farm income makes farming less relevant and therefore encourages the family to remain in the cooperative, where they receive lower farming income (see Table 1) but are exposed to less risk. The coefficient of the last physical capital variable – the number of animals – is negative and statistically significant. This is somewhat surprising, since one would expect people with more cattle to prefer independent farming. Yet, the availability of animals makes people less likely to leave the cooperative, despite their greater capital endowment. This can be explained by the fact that membership in the cooperative provides easy access to collective pastures at a relatively low cost. Buying cattle was indeed one of the main strategies for APC members to accumulate individual wealth, making use of free grazing rights at collective rangelands. Our regression model is based on the underlying assumption that the decision to leave the cooperative in favor of independent farming is mainly determined by the human and physical capital endowments of the household. The results produced by the model suggest, however, that this decision is also influenced by other institutional factors that are not included explicitly in the model specification. The field survey provides some indications regarding the main motives for staying in the cooperative. Table 5 presents the frequency distribution of the reasons given by group III respondents (those who remain in a cooperative). Income earned in the cooperative is
16
given as the main reason by only 5% of the respondents and it is thus a marginal factor in the decision to stay. One third of the respondents attribute their decision to stay to the perceived advantages of cooperation and joint action: 12% indicate that they like working jointly with others and 21% see distinct advantages in the provision of cooperative services. Yet more than 60% of the respondents explain their choice to remain in a cooperative by institutional and organizational constraints dating back to the creation of APCs under the Sandinista regime. Fully 43% of respondents remain in a cooperative because of the uncertainty associated with the ownership of land that had been expropriated from the large estates and distributed to the population by the Sandinista government in the 1980s. Another 19% do not leave because of lack of legal arrangements for the resolution of the outstanding debt burden, which was accumulated as a result of the generous credit support available to the cooperatives during the Sandinista era.11
Table 5. Reasons given by respondents for remaining in a cooperative Reasons Percent of group III respondents (N=167) Income in the cooperative 5 Likes working jointly with others 12 Access to cooperative services 21 Uncertainties with land ownership 43 Lack of mechanisms for resolution of cooperative debt 19
While capital endowments may be important for peasants’ decisions, it seems from these qualitative questions that two institutional constraints – uncertainty associated with land ownership and difficulties with resolution of cooperative debt – play a dominant role in keeping Nicaraguan peasants in APCs. The respondents are sending a very clear signal that, until these two constraints are resolved, peasants are actually forced to remain in cooperatives. Those who choose to leave under the present circumstances ultimately may face the danger of losing their land when ownership rights are cleared and legalized or may find themselves faced with an insupportable obligation representing their share of cooperative debt.
17
Reasons to Stay: The Role of Social Capital Access to cooperative services and options for participating in joint activities are frequently mentioned in the literature as relevant reasons for remaining a member in a cooperative. Social capital clearly plays an important role as a mechanism for sharing risk with others and as a device for maintaining access to service provision. Access to services has become a particularly important consideration in Nicaragua, since state support to agriculture was substantially reduced after the collapse of the Sandinista government. Privatization of the rural banking system and decentralization of extension services (mostly without appropriate budgets) force peasant households to look for alternative institutional relations. Local networks with private agents (traders, moneylenders) and linkages with the voluntary sector (NGOs and externally financed rural development projects) are now of primary importance for the peasantry in guaranteeing access to credit, market outlets, inputs, and information. To analyze the importance of social capital, we considered the role of cooperative relationships as a factor contributing to household expenditures or enabling access to rural financial markets. The survey database was used to determine whether peasants who are more involved in cooperative activities are actually better off in terms of welfare and borrowing options. We use a broad definition of cooperation, including participation in various types of social and institutional networks. Attention has therefore been given to community ties, various types of exchange relations (joint input provision and marketing), and group production and service delivery as indicators of the available social capital. Even when peasants decide to withdraw their land from the cooperative, they can still maintain some involvement in other networks with former members building on earlier cooperative experience.
18
The empirical analysis addresses the significance of social capital for the welfare of peasant households. Attention is given to the direct contribution of social capital to the level of household expenditures and the indirect role of social capital as additional collateral for borrowing. We used OLS and Probit regressions to identify the impact of physical, human, and social capital on household expenditures and on access to credit, respectively. Monthly expenditure is a reliable indicator of a household’s permanent income. All three groups of farmers were used in the expenditure regression, where the dependent variable included expenditures on food, clothes, education, transport, medicines, and electricity. Social capital, physical capital, and institutional networks are used as major explanatory variables for variations in expenditure, together with relevant individual and household characteristics (age, gender, education and family size). For the Probit regression with access to credit as the dependent variable we constructed a random balanced subsample of farmers with and without access to financial services to guarantee normality. Access to credit is registered as the possibility of using financial services provided by formal institutions (banks) or semi-formal agencies (NGOs, moneylenders). Access may be less than actual use, since some peasants may prefer not to borrow, either due to risk aversion or because internal financial resources are available. About two-thirds of the peasants in our sample declared to have access to credit, whereas only 45% actually took loans. Most credit was provided by NGOs and development programs (35%), followed by banks (27%) and friends or relatives (16%). Households with access to credit attained expenditure levels that were on average about 20% higher than for households without access to credit. Social and physical capitals were constructed as index variables by factor analysis,12 which enabled us to include relationships between different aspects and find their common
19
underlying dimensions. For social capital, the following variables were used: the number of connections of the household with various institutions (e.g., farmers unions, traders, moneylenders, banks, village authorities), the level of participation in cooperative organizations, the number of activities coordinated with other peasants (e.g., shared input purchase, marketing, joint machinery services), and the percentage of crop output marketed jointly with others. In a similar way, physical capital is calculated as an index variable that includes the number of capital goods (tools, tractors, implements, warehouses), the number of luxury goods (radio, television, motorbike, car), the number of large livestock (cows and horses), and the amount of land owned by the household (in manzanas). In addition, institutional linkages of households with market agents and the voluntary sector were included in the analysis (the state sector was ignored, since only 17 % of the sample still maintain connections with this sector). We used a dummy variable to indicate whether the household was primarily oriented towards market agencies (traders, moneylenders, shops) or voluntary sector organizations (NGOs and development projects).
Table 6. Impact of social capital on expenditures and access to credit Expenditures (OLS regression) Dependent variable: Parameter Significance level Estimate Explanatory variables: Constant 7.296 0.000 Social capital (index) 0.239 0.000 Physical capital (index) 0.180 0.000 Education (years) 0.066 0.000 Household size (persons) 0.093 0.000 Age (years) 0.035 0.017 Age squared -0.001 0.054 Gender (0=male; 1= female) 0.069 0.527 Distance to market (km) -0.002 0.510 Sector (0= voluntary; 1=market) -0.031 0.579 R2 0.38 Access to credit (Probit) Parameter Significance level estimate 1.157 0.244 0.699 0.000 -0.162 0.066 0.018 0.594 0.017 0.647 -0.042 0.266 0.000 0.286 0.720 0.471 0.027 0.023 -0.545 0.006 0.21
Social, physical, and human capitals have a positive and significant impact on household expenditures (Table 6, first two columns). Moreover, expenditures increase with age, but tend to decrease at higher ages, illustrating the nonlinear effect discussed before. The relation between
20
household size and expenditure is also positive, indicating that additional family members offer a positive contribution to household income. Social capital is particularly important for increasing the probability of access to credit for rural households, while physical capital surprisingly appears with a negative sign (see Table 6, last two columns). The latter result is surprising because one would expect that physical capital serves as collateral for borrowing. A possible explanation can be found in the highly insecure legal environment in Nicaragua, where registration of property rights is inadequate and many conflicting claims exist (Strasma 1998). As a result, the main components of physical property – land and buildings – are seldom used as collateral, since these assets cannot be sold in case of default. Therefore, social capital tends to be the preferred collateral, enabling group members to assume joint responsibility for the credit contract. Distance appears to have an unusual positive impact on access to credit. This could be explained by the fact that most farmers nowadays obtain credit though local NGOs and rural development projects. Donors generally prefer geographical targeting to poor people that are likely to be located in more remote and marginal regions. This is consistent with the fact that access to credit improves when a household maintains primarily relations with voluntary sector agencies. No indications were found for gender bias in credit provision. Most credit programs give a high weight to poverty targeting and the existence of local organizations, but tend to disregard the strong male dominance in traditional societies (Deere and Leon 2001). The social capital analysis reveals the positive effect of participation in social networks on household welfare and access to finance. The direct benefits of cooperative membership as a device for reducing transaction costs may be a reason for the slow transition towards independent farming. In addition, cooperative members also maintain some degree of coordination of
21
activities in order to capture the benefits of international cooperation projects and programs. The latter strategy is reinforced by the prevailing preference of foreign agencies to finance collective activities and the neglect of national agencies to provide substantial support for the development of independent farming in more remote areas. Concluding Remarks Production cooperatives have been established in Nicaragua during the Sandinista land reform process of the 1980s. Landless families were encouraged to join cooperatives rather than request a plot of land for individual farming. After about a decade, the socialist-oriented Sandinista regime was replaced in elections by a market-oriented government. Liberalization of the economy implied that former preferential treatment of agricultural cooperatives was abandoned. Whereas some of the cooperatives split up into individual farms, not all peasants abandoned cooperative membership. We analyzed differences in income position, resource use, and living standards between former cooperative members, peasants who continued their cooperative membership, and individual peasants who were already engaged in independent farming before the land reform. Rural families that leave the cooperative usually remain with limited land and capital resources. Peasants who stay in the cooperative do so because of major uncertainties regarding land ownership and outstanding debts. In addition, they benefit from better access to services and use their social capital to gain access to credit. Nicaraguan policies towards privatization of land ownership was ambiguous in the sense that no additional conditions have been created to enable farmers who decided to leave the cooperatives to overcome accumulation barriers. One of the main conclusions of this study is a reaffirmation of a fact that has been frequently observed in many developing and transition economies: access to land is the main
22
determinant of well being in rural communities. Giving rural people more land is the surest way of alleviating rural poverty, increasing incomes, and improving family welfare. It is therefore particularly important to eliminate the existing constraints to effective distribution of land to individuals in rural Nicaragua. People who leave the cooperative should be given their full and fair share of the land as a precondition for starting successful independent farming. The survey seems to indicate that former cooperative members are “short-changed” when they leave the cooperative: they get less than their fair share of land and assets and are thus handicapped from the outset in the new and risky endeavor of private farming. To enable peasants to make a free choice between cooperative membership and individual farming, the government must deal with two institutional factors that have been clearly identified by the respondents as major obstacles to exit from cooperatives. First, it is essential to resolve the uncertainty regarding land ownership. APC members must be assured that the land they have been using for more than two decades will remain theirs regardless of the ownership antecedents. Where compensation is a relevant issue, former owners should be compensated by the state with money or monetary instruments, not through restitution. It is inconceivable from considerations of social justice that, after more than twenty years, the poverty-stricken peasants in Nicaragua should face the danger of losing their land – the main source of livelihood – through restitution to former estate owners. In the same vein, it is necessary to eliminate the uncertainty and opacity concerning the disposition of old debts accumulated by the cooperatives during the Sandinista period. The cooperative members cannot be held responsible for the creation of this debt: it was thrust upon them by government policies and it should now be lifted again by government decision. As in many other countries, writing off this old debt is probably the only socially just option if the
23
government wishes to enable the rural population to make their own choice between remaining cooperative members or starting up independent farming. Finally, the Nicaraguan government and (inter)national organizations could contribute substantially to creating a more equitable environment for the peasantry through the provision of legal assistance for settling land conflicts and the delivery of rural financial services – including credit, savings, and insurance – for enabling farmers to improve their efficiency and to consolidate their landholdings. Especially in remote regions, where markets are less integrated and traditional patron–client relations still prevail, leaving the cooperative to take up individual farming involves severe institutional constraints. In this setting, there is undoubtedly room for further supportive measures aimed at reducing the transaction costs and the risks of independent farming. Bibliography Amelina, M. (2000) ‘Why Russian Peasants remain in Collective Farms: a Household Perspective on Agricultural Restructuring’, Post-Soviet Geography and Economics Vol. 41, pp 483-511. Baumeister, E. (1998) Estructura y Reforma Agraria en Nicaragua 1979-1989. Managua: CDRULA. Bardhan, P. Ed.(1989) The Economic Theory of Agrarian Institutions. Oxford: Clarendon Press. Barham, B. and M. Childress (1992) ‘Membership Desertion as an Adjustment Process of Honduran Agricultural Reform Enterprises’, Economic Development and Cultural Change, Vol. 40, pp. 587-614. Bromley, D. (1992) Making the Commons Work: Theory, Practice and Policy. San Francisco: ICS Press. Brooks, K.M. (1993) Decollectivization and the Agricultural Transition in Eastern and Central Europa. Working Paper WPS 993. Washington DC: World Bank.
24
Carter, M., J. Melmed-Sanjak and K. A. Luz (1993) ‘Can Production Cooperatives resolve the Conundrum of Exclusionary Growth? An Econometric Evaluation of Land Reform Cooperatives in Honduras and Nicaragua’, in: C. Csaki and Y. Kislev (eds) Agricultural Cooperatives in Transition. Boulder CO: Westview Press. Carter, M. (1987) ‘Risk Sharing and Incentives in the Decollectivization of Agriculture”, Oxford Economic Papers, Vol. 39, pp. 577-595 Deere, C.D. and M. Leon (2001) Empowering Women: Land and Property Rights in Latin America. Pittsburgh: University of Pittsburgh Press, pp. 62-106. Deininger, K. (1995) ‘Collective Agricultural Production: A Solution for Transition Countries?’, World Development, Vol. 23, pp. 1317-34. Deininger, K. (1993) Cooperatives and the Breakup of Large Mechanized Farms: Theoretical Perspectives and Empirical Evidence. Discussion Paper No. 218. Washington DC: World Bank. Everingham, M. (2001) ‘Agricultural Property Rights and Political Change in Nicaragua’, Latin American Politics and Society, Vol.43, pp. 61-93. Hazell, P. (2003) ‘Ensuring a Future for Small Farms in the Developing World’, Keynote Address 25th International Conference of Agricultural Economists, Durban, August 2003. Hoff, K., A. Braverman and J.E. Stiglitz, Eds. (1993) The Economics of Rural Organization. Oxford: Oxford University Press. Jonakin, J. (1992) Nicaraguan Collective Cooperatives: A Study of Choice, Efficiency and Performance. Doctoral dissertation, University of Tennessee. Kennedy, L. (1983) The Economic Theory of Cooperative Enterprises. Oxford, Plunkett Foundation for Cooperative Studies. Lerman, Z., C. Csaki, and G. Feder (2004). Agriculture in Transition: Land Policies and Evolving Farm Structures in Post-Soviet Countries, Lanham MD: Lexington Books. Putterman, L. (1985) ‘Extrinsic and Intrinsic Problems of Agricultural Cooperation: Antiincentivism in Tanzania and China’, Journal of Development Studies, Vol. 21, pp. 175204. Putterman, L. (1981) ‘On Optimality in Collective Institutional Choice’, Journal of Comparative Economics, Vol. 5, pp. 392-403. Ruben R. and E. Masset (2003) ‘Land Markets, Risk and Distress Sales in Nicaragua: the Impact of Income Shocks on Rural Differentiation’. Journal of Agrarian Change, Vol. 3, No. 4, pp. 481-199.
25
Ruben, R. (1999) Making Cooperatives Work: Resource Management and Contract Choice within Agrarian Reform Cooperatives in Honduras. Amsterdam: CEDLA-CLAS. Ruben, R. and F. Funez (1993) La Compra-Venta de Tierras de la Reforma Agraria. Tegucigalpa, Honduras: Guaymuras. San Martin, O. (1992) El cooperativismo Centroamericano en Cifras. San José, Costa Rica: CCCCA. Strasma, J. (1998) Land Tenure in Nicaragua: Analysis and Future Perspectives. Rome: FAO Land Tenure Service. Wilson, P. and G. Thompson (1993) ‘Common Property and Uncertainty: Competing Coalitions by Mexico’s Pastoral Ejiditarios’ , Economic Development and Cultural Change, Vol. 41, pp. 299-317. World Bank (2003) Agriculture in Nicaragua: Promoting Competitiveness and Stimulating Broad-Based Growth. Washington DC: The World Bank.
26
NOTES
Field research for this paper was supported by the Netherlands–Israel Research Program (NIRP) grant 96-12.1. The project was carried out in collaboration with the School of Agricultural Economics (ESECA) at the National University of Nicaragua in Managua. Gustavo Siles, Orlando Cortes, Luis Rodriguez, Estela Aleman, Yani Jarquin, and Irena Guevara supervised the field work in Nicaragua. Support for data analysis was provided by Jos Vaessen, Danielle van Strien, Edoardo Masset (Wageningen University), Val Khutemliansky, Yulia Rabinovich, and Daria Zaslaver (The Hebrew University). Socialist agriculture included both collective and state farms, but it is collective farms (or cooperatives, as they were called in Central Eastern Europe) that dominated the rural space in the entire region. With the start of market reforms in the early 1990s, all state farms in the former Soviet Union were transformed into collective farms and their land and assets were privatized to the employees and the pensioners. In Central Eastern Europe, state farms were generally sold as going concern to outside investors or allowed to go bankrupt. Therefore, in our context, the relevant issue is the persistence of the large contingent of collective farms, which were quickly transformed after 1990 into membership-based corporate structures in all transition countries. For details on differences between collective and state farms and their restructuring procedures see Lerman et al. (2004). Random sampling took place in four different agro-ecological zones defined according to major cropping systems. Data for the analysis were collected by a two-stage (village/farmer) sampling procedure. The sampling frame used for the random selection of villages and farmers was obtained from the Ministry of Agriculture and FENACOOP (National Cooperative Federation) listings. Fieldwork concentrated in the semi-arid León-Chinandega with traditional agro-export crops (cotton and sugar production), the semi-humid southern region comprising the Carazo mountains (coffee) and the Masaya-Granada plain area (horticulture crops and rice production), the mountainous interior area of Boaco and Chontales (livestock and cereal production), and the northern region of Matagalpa-Estelí (coffee production). In each of these four macro-regions, a random sample of 120 farm households (valid cases) was used, including equal proportions of members of agricultural production cooperatives, former members who are now engaged in independent farming, and small and medium-size independent peasants.
4 3 2 1
We selected independent farmers with a farm size up to 40 manzanas (28 hectares) as a base for comparison, in line with the average plot size of cooperative members.
This paper only analyzes the peasant’s decision to exit the cooperative and continue farming independently. The other alternative of leaving agriculture altogether and switching to a non-farming occupation (including migration to the city) is not included in this analysis. This alternative is analyzed by Barham and Childress (1992). Sale and breakdown of land reform cooperatives in Honduras is analyzed by Ruben and Funez (1993). State farms operated in Nicaragua until the mid-1980s but were strongly reduced at the end of that decade. The Sandinista government originally favoured state production, but had to abandon this policy after fierce peasant resistance. At the time of our survey, state farms represented less than 4% of land use (Strasma 1998).
7 8 9 6
5
This also includes a small number of ‘traditional’ cooperatives created before the Sandinsta era. See for a concise overview of these issues: Deininger (1993).
The low explanatory power of the model (R2 = 0.16, see Table 2) is due to large noise in the data and omitted variables in the model.
10
The survey questions explicitly distinguished between farm assets owned by the cooperative farmer as an individual and by the cooperative as an organization.
11
Credit to agriculture (as a share of agricultural GDP) declined from from 43 in the early 1990s to only 14% by the end of 2001 (World Bank 2003). Detailed results of the factor analysis can be obtained from the authors.
12
27